Hugo is racing down College Street on an uncomfortably hot afternoon, late for a meeting with students. He skipped lunch and is now experiencing hunger pangs. Suddenly, he spies a small unattended child, whose mother has just popped into the post office, enjoying a voluminous ice cream sundae. He experiences a strong urge to snatch the sundae from the unsuspecting child and eat it himself. However, almost simultaneously, Hugo judges that it would be morally wrong for him to do that. He experiences this moral judgment as exerting psychic pressure on him to abstain from stealing the sundae. After a momentary internal struggle, Hugo’s better judgment wins out: He shoots past the child and on to his meeting.

As the vignette illustrates, it is part of our commonsense conception of the mind that there is an intimate connection between moral judgment and motivation: Such judgments seem to exert motivational pressure on their host subject to act as they recommend. In this chapter, we critically review the literature on moral motivation. Philosophers and psychologists alike have been puzzled by the phenomenon of moral motivation: How can a moral judgment exert motivational force? The nature of moral motivation has long been seen as a “hinge issue” around which core debates in moral philosophy – over the nature of moral judgment and the reality of objective moral facts, etc. – turn. Here we sketch the contours of this pivotal issue in moral psychology and catalogue and define the various key theoretical positions. We then turn our attention to evaluating the relevance of certain empirical results that, some philosophers have argued, bear on these positions.

3.1 What Are Motivational States?

Before we investigate the nature of moral motivation, we ought to first say something about the nature of motivation in general. Paradigm examples of motivational states include desires, intentions, and emotions such as fear, etc. But what unites these mental states together as distinctively motivational? What feature do they have in common, in virtue of which they all count as being a motivational state and that explains their motivational character?

We propose to (partially) answer this question by reviewing some of the folk-psychological commitments that are broadly agreed by philosophers to characterize the class of motivational mental states. The first such commitment is that motivational states are intimately related to action – yet distinct from it. If you intentionally perform action F, then you must have had some motivation to F. But this motivation to F, by itself, is insufficient for your actually performing action F. After all, you can be motivated to F yet fail to do F for a variety of reasons – for example, because you were more strongly motivated to perform action G that was incompatible with your doing F.

Second, motivational mental states often appear to produce action only in tandem with the right background beliefs (Smith, Reference Smith1994). My desire to drink soda, in tandem with my belief that I can drink soda if I walk to the soda machine and insert a dollar into it, can prompt me to do just that. Likewise, my intention to drink from the bottle of soda I am holding, together with my belief that I can drink from that bottle if I raise it to my lips, can cause me to raise the bottle to my lips. As these cases suggest, a background belief (of the right sort) appears to be necessary for a motivational mental state to produce action. If I had no beliefs about how I might acquire soda, then my desire for soda wouldn’t prompt me to perform any particular action. And such beliefs play an important role in determining which action a motivational state will bring about. After all, had I believed that I could acquire soda by praying to the Carbonated Beverage Gods, then my desire for soda would have caused me to feverishly pray rather than walk to the soda machine. In general, then, motivational mental states appear to occupy a certain (coarse-grained) functional role in your mental economy: They can produce action in conjunction with the right (means–end) belief.Footnote ¹

Third, motivational mental states play a particular functional role, distinct from that played by means–end beliefs, in the production of action: In particular, they set the target for action and provide the impetus, or “push,” that generates it. In contrast, means–end beliefs, although necessary for the generation of action, play a mere “coordinating” or “guiding” role: They pick out the causal relations among various actions, that relate means to ends (Dretske, Reference Dretske1988). Such beliefs are not the driving source of action; that is the role of the motivational states. Rather, their role is to guide motivational states toward a successful realization of their aim. So, although motivational mental states and means–end beliefs are each necessary and jointly sufficient (in the right circumstances) for the production of action, they play very different roles in the psychic generation of overt behavior. The former provides “the aim and the impetus,” the latter the guidance.

Fourth, it is widely held that motivational states are subject to different norms than those governing beliefs and other cognitive mental states. Belief, it is broadly agreed, is governed by epistemic norms alone (Adler, Reference Adler2002; Parfit, Reference Parfit2011; Shah, Reference Shah2006; Way, Reference Way2016). This is the doctrine known as evidentialism. Epistemic norms include requirements of theoretical rationality – such as the prohibition against believing contradictory propositions – and considerations that count as evidence in favor of one proposition or another. So the only factors relevant to whether or not you should believe the proposition that the Moon is made of cheese are epistemic norms – such as our decisive evidence from scientific inquiry that the Moon is made only of noncheese substances. Practical considerations, such as your self-interest, are neither here nor there. For example, the fact that an eccentric billionaire has promised you ten million dollars if you believe by next Tuesday that the Moon is made of cheese is not a reason for you to believe that proposition – although it could certainly be a reason for you to take steps to ensure that you believe by next Tuesday that the Moon is made of cheese, perhaps through experimental neurosurgery or hypothetical “belief pills.” Of course, there are dissenters from this evidentialist orthodoxy: Pragmatists about belief hold that practical considerations can count (sometimes) as reasons for belief (Hieronymi, Reference Hieronymi2005; Leary, Reference Leary2017a). But the standard evidentialist view has it that belief is subject to epistemic norms alone.

Now, whereas beliefs and other cognitive states are governed by epistemic norms alone, motivational mental states are also governed by distinctively practical norms. By way of illustration, suppose that I have much to gain from ingratiating myself to the host of the party I am attending tonight. Given this, I have strong reason to ingratiate myself to her – perhaps through engaging her in charming small talk. Consequently, I am warranted in forming the intention to attempt to ingratiate myself to my host. This case demonstrates how intention, unlike belief, is governed by practical considerations: In light of my strong reasons to F, I am justified in forming an intention to (attempt to) F.

It is also highly plausible that intention is governed, in addition, by certain epistemic norms. Suppose, for example, that I am trapped down a well. If I had good evidence that I could take flight by flapping my arms, then it would be rational for me to form the intention to do so. After all, I could remove myself from my current predicament by flapping my arms. However, very plausibly, you cannot rationally intend to do something that you cannot rationally believe you can possibly do. Given my evidence, I cannot rationally believe that I can possibly take flight by flapping my arms. Consequently, I cannot rationally intend, given my evidence, to so take flight. Thus, your intentions are governed by an epistemic norm of consistency with your evidence, in addition to distinctively practical norms concerning what you have reason to do (Bratman, Reference Bratman1987; Holton, Reference Holton2009; Marusic & Schwenkler, Reference Marusic and Schwenkler2018; Ross, Reference Ross2009; Setiya, Reference Setiya2008; Velleman, Reference Velleman1989).

Motivational mental states beyond intention, such as desire, are also – according to some philosophers (Parfit, Reference Parfit2011; Scanlon, Reference Scanlon2013) – governed by practical norms. For example, in light of our decisive moral reasons to avoid killing others with our actions, we ought to form the desire to avoid killing others. This is how a fully rational agent, it is claimed, would respond to these reasons. Of course, there are dissenters from this view. For example, those who endorse a Humean view of reasons hold that our (ultimate) desires are not subject to reason or rational criticism. Rather, our (ultimate) desires are the source of our reasons for action. You have a reason to F, for the Humean, just when you have some ultimate desire that would be satisfied (at least, somewhat) if you performed action F (Schroeder, Reference Schroeder2007). Nevertheless, on both views, there is an intimate relation between reasons for action and desires to so act. This contrasts with the case of belief and other nonmotivational cognitive mental states. These mental states, by the lights of common sense (Stahl et al., Reference Stahl, Zaal and Skitka2016), are subject to epistemic norms alone. But motivational mental states, as we have just seen, should be understood too through reference to their connections to distinctively practical norms. So, it is not just their characteristic functional role that distinguishes the motivational mental states from the nonmotivational cognitive ones, but also the nature of the norms to which they are subject or related. This is a key aspect of our folk-psychological understanding of the motivational, and the mental more broadly.

Let us summarize our theoretical position on the nature of the motivational mental states. Motivational mental states can produce action in tandem with the right means–end beliefs; play a certain functional role in the production of action (they provide “the aim and the impetus”); and are governed by distinctively practical norms (in addition, perhaps, to epistemic norms). Of course, the functional roles constitutive of, and norms governing, particular motivational mental states – the desire that p, the intention to F, etc. – can be articulated in a more fine-grained manner. For example, the functional role constitutive of desire might be thought to be fully expressed by the role assigned to it by normative decision theory (Lewis, Reference Lewis1988, Reference Lewis1996) or a completed cognitive neuroscience (Schroeder, Reference Schroeder2004). And the functional and normative nature of intention has also been more finely articulated (e.g., Bratman, Reference Bratman1987). But the features we have catalogued here capture, we believe, what is distinctive of motivational states qua motivational states.

3.2 The Problem of Moral Motivation

Let us now turn to the nature of moral motivation. The locus classicus of the debate over the nature of moral motivation is a line of thought widely attributed to David Hume in his (1739) Treatise of Human Nature.Footnote ² We propose – taking inspiration from Smith (Reference Smith1994) – to formulate this dialectic as an inconsistent triad (i.e., a trio of jointly inconsistent propositions) that we dub “Hume’s problem.” Despite being jointly inconsistent, each of the three propositions is found highly plausible by philosophers. Each expresses a central organizing principle in moral philosophy. Moreover, each proposition can be empirically investigated. The three doctrines in question are the Humean theory of motivation; cognitivism about moral judgment; and moral judgment internalism. But, given their joint inconsistency, one must be given up. And the choice you make here determines, to a significant extent, the theoretical options available to you in a host of central debates in moral philosophy.

Hume’s problem is formulated as follows:

1. (1) Beliefs cannot causally suffice, by themselves, for motivation.

2. (2) Moral judgments are beliefs.

3. (3) Moral judgments causally suffice, by themselves, for motivation.

Proposition (1) is the statement of the Humean theory of motivation. Proposition (2) expresses cognitivism about moral judgment. And proposition (3) articulates moral judgment internalism. Before we outline the inconsistency Hume finds between these doctrines, let us first characterize each of these views and review the reasons for endorsing them.

The Humean theory of motivation is the doctrine that beliefs alone cannot causally suffice for motivation. Rather, desire is necessary too: Motivation and action are generated only by a belief and a desire working in tandem (Schueler, Reference Schueler2009; Sinhababu, Reference Sinhababu2009; Smith, Reference Smith1987, Reference Smith1994). Desires can be brought about, or changed, by some chain of reasoning only if a desire features among the premises of that thinking (Sinhababu, Reference Sinhababu2009). Very briefly, this principle of philosophical psychology enjoys the status of orthodoxy on the grounds that it seems to be part of our (empirically very successful) folk-psychological theory of the mind: Our folk-psychological grip on the mental domain suggests that beliefs and motivational states bear no logically necessary connections to one another. After all, Hume noted, it seems that having any one set of beliefs regarding the ways things are, rather than some other set of beliefs, doesn’t in itself place any restrictions on how one is fundamentally motivated to act. Beliefs by themselves don’t appear to cause, or necessitate, particular desires or intentions, etc. Of course, my desire for a thirst-quenching drink might combine with my belief that I can get a drink from the refrigerator to cause me to acquire the intention to walk to the fridge. But this is consistent with beliefs alone having no causal or necessary connection to desires or to other elements of motivation. And this latter claim is all the Humean is affirming. In this way then, our commonsense grasp of the mental supports the Humean’s contention that beliefs cannot causally suffice, by themselves, for motivation. Hence, in advance of our theoretical commitments pushing us this way or that, we ought to endorse the Humean theory of motivation. It enjoys enough prima facie warrant to constitute the presumptive view in this region of philosophical psychology.

The second doctrine making up “Hume’s problem” is cognitivism about moral judgment. This is simply the view that moral judgments are beliefs – in particular, beliefs whose contents are propositions concerning moral requirements or reasons. Examples of such judgments are my belief that it would be wrong for me to eat meat or your belief that you are morally required to give to charity.Footnote ³ Cognitivism contrasts with noncognitivism about moral judgment. This is the doctrine that moral judgments are not beliefs, but rather certain (complexes of) noncognitive mental states, such as desires, sentiments, or states of approbation or disapprobation, etc. The phenomenon of morality, for the noncognitivist, bottoms out then, not in a realm of moral facts, but rather in the fact that we happen to be motivated to behave in certain – pro-social; transgressor-punishing; norm-endorsing, etc. – ways (Blackburn, Reference Blackburn1998; Gibbard, Reference Gibbard1990).

The central importance of the debate between the cognitivist and the noncognitivist to moral philosophy can be appreciated when it is observed that cognitivism is a commitment of moral realism, the view that there are (objective) facts about what morality requires of us. After all, if there are moral facts, then moral judgments must be the mental states that aim to correctly represent these facts. Since beliefs are the kind of mental state that aims to correctly represent – or fit – the facts (Anscombe, Reference Anscombe1957; Smith, Reference Smith1987, Reference Smith1994), the moral realist must hold that moral judgments are a variety of belief. To judge that morality requires you to F, is just to believe that morality requires you to F. Conversely, the truth of noncognitivism seems to entail the truth of moral antirealism, the doctrine that there are no facts of the matter about what morality requires of us. This opens the way for a potent antirealist argumentative strategy: simply show that cognitivism is false and, from there, derive the truth of moral antirealism (Ayer, Reference Ayer1936).

The third, and final, doctrine making up Hume’s inconsistent triad is moral judgment internalism. This is the view that moral judgments can causally suffice, by themselves, for motivation (Prinz, Reference Prinz, Strandberg, Björklund, Björnsson, Eriksson and Olinder2015). More precisely, it is the doctrine that your judgment that morality requires you to do F causally suffices, by itself, for your being motivated – to some degree – to do F. In other words, that your moral judgment here can incline or push you – somewhat, at least – toward doing F. It contrasts with moral judgment externalism, the view that your judgment that morality requires you to do F does not causally suffice, by itself, for the presence of any motivation in you toward doing F. Rather, moral motivation can only result from moral judgment in conjunction with some background mental state. Philosophers who endorse internalism include Blackburn (Reference Blackburn1998), Copp (Reference Copp2018), Darwall (Reference Darwall1983), Dreier (Reference Dreier, Smith and Johnson2015), Gibbard (Reference Gibbard1990), Mackie (Reference Mackie1977), and Smith (Reference Smith1994).

Moral judgment internalism looks to enjoy intuitive support. Arguably, the first-person phenomenology of making a moral judgment suggests that such judgments can causally suffice, by themselves, for the presence of motivations to act as they recommend. Suppose, for example, that I am in dire financial straits: An investment has gone sour and I am overleveraged on my mortgage. While filling out my taxes for the year, I realize I could salvage my financial situation somewhat by misrepresenting my income to the Federal Government. I would likely get away with it, I muse, and the negative impact on others would be negligible. I then experience the urge to cheat on my taxes. However, I suppress this urge, in part, by reminding myself that it would be morally wrong for me to cheat on my taxes. Introspectively, it certainly seems like my first-person moral judgments are always accompanied by some motivation in me to act in accord with them. I experience them as exerting psychic pressure on me to act as they recommend.

In addition, moral judgments are often cited as playing the role of a motivational state in the explanation of action. For example, suppose that I ask an eminent historian why certain Polish partisans sheltered Jews during World War II. She replies: “Because they judged that morality required them to save the lives of the innocent and they believed that they could save the lives of these innocent people by sheltering them.” Here moral judgments are depicted as playing the role of a motivational mental state in the production of action: In tandem with certain means–end beliefs, moral judgments causally suffice for action. Examples such as these are widely agreed to provide prima facie support for internalism (Shafer-Landau, Reference Shafer-Landau2003).

The internalist should say more about what precisely she means when she says that a moral judgment can causally suffice, by itself, for motivation. Some contemporary internalists flesh their doctrine out in the following way: Internalism, they maintain, is the view that moral judgments causally suffice, by themselves, for motivation in a rational agent. This is a version of “conditional internalism,” internalism defeasible under certain conditions, that is currently the most prominent variety of internalism endorsed in the contemporary philosophical literature (Korsgaard, Reference Korsgaard1986, Reference Korsgaard1996; Smith, Reference Smith1994; van Roojen, Reference van Roojen and Zalta2018; Wallace, Reference Wallace and Dreier2006; Wedgwood, Reference Wedgwood2007). It is a step back from unconditional internalism, the view that moral judgments are necessarily motivating, that characterized the earlier literature on internalism. This new weaker version of internalism was motivated by the recognition that being in a psychologically abnormal condition, such as apathy or depression, one incompatible with full rationality, can render one’s moral judgments motivationally inert.Footnote ⁴ Externalism should now be understood, in contrast, as the doctrine that moral judgments do not causally suffice, by themselves, for motivation in a rational agent. Rather, if moral judgments do causally suffice for appropriate motivation in a rational agent, then it must be in virtue of said judgments interacting with some or other background attitude(s).

We are now in a position to understand Hume’s problem and its force. The alleged inconsistency between the Humean theory of motivation, cognitivism, and internalism can be rationally reconstructed in the following manner: Suppose that you judge that you are morally required to F. Granting (rationalist conditional) internalism – the doctrine that your moral judgments causally suffice, by themselves, for the presence of motivation to act in accord with these judgments in a rational agent – it follows that you must be in a motivational mental state that can incline, or push, you toward doing F. But the Humean theory of motivation has it that no belief can causally suffice, by itself, for the presence of any motivation whatsoever, even under conditions of full rationality. This means that your moral judgment cannot be a belief, and must rather itself be a motivational state of one sort or another – most plausibly, some kind of desire. In this way then, we have deduced the falsity of cognitivism about moral judgment, the doctrine that moral judgments are beliefs, and the truth of noncognitivism, from the conjunction of internalism and the Humean theory of motivation.

Of course, a stout-hearted cognitivist will not just roll over in defeat here. Rather, she will resist this dialectic by rejecting either internalism or the Humean theory of motivation (or both). The first group of cognitivists who reject internalism are externalists. For externalists, moral judgments alone do not causally suffice for motivation. Moral judgments are, by themselves, motivationally inert. Merely judging that you are morally required to do F does not, by itself, provide any motive to do F. Rather, moral judgments can only bring about moral motivation in conjunction with the right background desire – the desire, say, to do the morally right thing, or the desire to avoid harm, provided that avoiding harm is in that instance the morally right thing to do, etc. (Boyd, Reference Boyd and Sayre-McCord1988; Parfit, Reference Parfit2011; Railton, Reference Railton1986).

The second group of cognitivists, those who deny the Humean theory of motivation, are known as “anti-Humeans.” They hold that beliefs alone can produce motivation. It is not the case, for anti-Humeans, that motivation is only ever brought about by a belief if it is working in tandem with an appropriate desire. Consequently, on this view, your moral judgments can causally suffice for motivation, and intentional action, all consistent with these moral judgments themselves being nothing but beliefs (McNaughton, Reference McNaughton1988; Nagel, Reference Nagel1970; Platts, Reference Platts1991; Shafer-Landau, Reference Shafer-Landau2003).

As we have illustrated, the nature of moral motivation is a “hinge issue” around which central debates in moral psychology and moral philosophy turn. Philosophers have long jockeyed their intuitions for and against the doctrines of internalism, cognitivism, and the Humean theory of motivation, hoping to establish their preferred resolution to Hume’s problem. We now turn to reviewing and evaluating claims made by various philosophers that certain empirical observations might bear upon these matters.

3.3 Empirical Psychology and Moral Motivation

Philosophers have recently begun arguing that certain empirical findings from the psychological and brain sciences tell in favor of one or other doctrine in philosophical moral psychology. In the rest of this chapter, we review and critically discuss a sample of these claims from the literature. We shall focus our attention first on claims that empirical results settle the debate between internalists and externalists (or, at least, militate heavily in favor of one over the other), before turning to similar claims regarding the disagreement between cognitivists and noncognitivists, and between Humeans and anti-Humeans about motivation.

Philosophers have argued that the phenomenon of clinical psychopathy is pertinent to the internalist/externalist debate (Schroeder et al., Reference Schroeder, Roskies, Nichols and Doris2010). Traditionally, internalists and externalists have coaxed our intuitions over the logical possibility of amoralists: hypothetical individuals who make moral judgments without being motivated in the slightest to act as they recommend, despite being seemingly rational (Brink, Reference Brink1997; Shafer-Landau, Reference Shafer-Landau2003). If amoralists are logically possible, then internalism stands refuted. After all, the mere possibility of an amoralist would entail that it is not the case that moral judgments, by their nature, enjoy a necessary conceptual connection to motivation in a rational agent. Externalists have therefore long appealed to the seeming possibility – or conceivability – of amoralists to bolster their view. However, such conceivability arguments are problematic, both because some doubt that the conceivability of some state of affairs p is good evidence that p is possible (Putnam, Reference Putnam and Putnam1975), and also because people’s intuition about what is conceivable differ. Given this, an actual example of a real-life amoralist would be far more convincing evidence of their possibility.

On first examination, psychopaths appear to be a case of real-life amoralists. Although such people are – by and large – cognitively normal, they manifest little guilt, empathy, or remorse for morally wrong actions. Psychopaths are often perfectly intelligent, seem rational, and appear able to make appropriate moral judgments about a wide range of cases. Nevertheless, they frequently have a history of chronic antisocial behavior behind them – including, but not limited to, lying, stealing, torturing, and killing. And, even more disturbingly, they can engage in these actions without emotional cost. In short, psychopaths look cognitively equipped to make appropriate moral judgments, but are seemingly indifferent to the deliverances of such judgments or the dreadful consequences of their actions. This looks like good empirical evidence for externalism (Kelly et al., Reference Kelly, Stich, Haley, Eng and Fessler2007).

However, internalists disagree. Some philosophers have argued that the existence of psychopaths poses no problem for internalism (Nichols, Reference Nichols2004; Prinz, Reference Prinz2007). How so? Well, these thinkers argue that psychopaths don’t really understand morality or even grasp moral concepts at all. As Prinz (Reference Prinz2007, p. 43) puts it: “Psychopaths seem to comprehend morality, but they really don’t. They use moral terms in a way that deviates strikingly from the way non-psychopaths use those terms. These deviations suggest that they do not possess moral concepts.” In other words, some internalists have sought to render the existence of psychopaths consistent with their view by denying that psychopaths really make moral judgments. After all, if psychopaths don’t even make moral judgments in the first place, then the fact that they seem not to be motivated to act in accord with the norms of commonsense morality – including norms that they profess to recognize – is no obstacle for internalism. In this way, the internalist can continue to hold that moral judgments can causally suffice, by themselves, for motivation (and do so suffice in a rational agent), while admitting the existence of individuals with the psychic and behavioral profile characteristic of psychopaths.

Of course, the position that a subset of intelligent adult humans don’t grasp moral concepts is a heavy lift. The internalist had better have some principled reason to advocate for it, independent of her desire to preserve her theory. Fortunately for the internalist, a reason is available. Nichols (Reference Nichols2004) and Prinz (Reference Prinz2007) argue that a grasp of moral concepts presupposes being able to distinguish the requirements of morality from the requirements of convention. After all, very plausibly, a grip on any arbitrary concept F requires being able to reliably distinguish instances of Fs from instances of non-Fs. You don’t really understand the meaning of a concept if you are systematically confused about its extension under conditions of full information.

These philosophers then appeal to empirical evidence that, in tandem with the reasoning described earlier, supports the conclusion that psychopaths don’t understand the moral/conventional distinction. First, Blair (Reference Blair1995) found that incarcerated adult psychopaths do not reliably distinguish moral from conventional wrongs. They did not treat moral and conventional wrongs significantly differently and, unlike a control group of nonpsychopathic prisoners, they tended to ignore the victim’s welfare when explaining why some action was morally wrong. Second, Blair (Reference Blair1997) administered the moral/conventional wrongs test on children with psychopathic tendencies. They found that these children, unlike control children, tended to treat all wrongs as merely conventional. Morality, for these children, seemed no different from etiquette or conventions about which side of the road one should drive upon. These results are all the more striking when one learns that healthy children have already begun to master the moral/conventional distinction by the time they are three years old (Nucci, Reference Nucci2001; Smetana, Reference Smetana1981; Turiel, Reference Turiel1983). Taken together, these results appear to strongly support the conclusion that psychopaths lack a (full or proper) understanding of moral concepts. Prinz (Reference Prinz2007, p. 44) sums up this picture of psychopathy as “psychopaths can give lip service to morality, but their comprehension is superficial at best.” If this is correct, then the argument for externalism from psychopathy can be defused, in the way described earlier, through appeal to the claim that psychopaths do not really make moral judgments, a proposition that is itself supported by reference to empirical evidence that psychopaths do not (properly) grasp moral concepts.

However, the empirical evidence concerning whether psychopaths understand the moral/conventional distinction is inconsistent. For example, Aharoni et al. (Reference Aharoni, Sinnott-Armstrong and Kiehl2014) found that psychopaths correctly distinguished moral from conventional transgressions. This study employed considerably more participants (139) than Blair’s 1995 study did (20–40), and its result also cohere with other similar studies (Aharoni et al., Reference Aharoni, Sinnott-Armstrong and Kiehl2012). Aharoni et al. (Reference Aharoni, Sinnott-Armstrong and Kiehl2014, p. 179) conclude thusly: “The observed pattern of results comports with the alternative view that psychopathic individuals ‘know right from wrong but don’t care.’” If these studies are correct, psychopaths continue to pose a problem for internalism.Footnote ⁵

Another group that seems unmoved by moral judgment yet understands moral concepts are patients with ventromedial prefrontal cortex (vmPFC) lesions. Some have argued that they are real-life cases of amoralists: Damasio et al. (Reference Damasio, Tranel and Damasio1990) describe the pattern of behavioral and psychic deficits manifested by such patients as constituting an “acquired sociopathy.” Roskies (Reference Roskies2003) argued that vmPFC patients constitute good empirical evidence for unconditional externalism about moral judgment and, because they possessed normal moral concepts prior to their injury, provide reason to believe that their cognitive grasp of morality remains intact. However, the moral judgments of vmPFC patients, according to Roskies, have lost their motivational punch. On Roskies’ analysis, the distinctive psychopathology of vmPFC patients reveals that moral judgments do not causally suffice, by themselves, for the presence of appropriate motivation. Disconnected from motivational systems by vmPFC lesions, their moral judgments fail to lead to motivation and action.

Let’s get clearer on the nature of the standard deficits induced by a vmPFC lesion. First, such patients appear cognitively normal on a wide range of standard psychological tests, including those measuring intelligence or domain-general reasoning abilities and those probing whether their knowledge of the world has been damaged. In particular, the moral reasoning of vmPFC patients appears to be unimpaired under experimental conditions: They perform at a normal level on Kohlberg’s moral reasoning scale (Saver & Damasio, Reference Saver and Damasio1991) and make normal moral judgments in a variety of hypothetical scenarios (Koenigs et al., Reference Koenigs, Young, Adolphs, Tranel, Cushman, Hauser and Damasio2007).Footnote ⁶

Second, vmPFC patients have a profound difficulty in acting in accord with social norms or, indeed, in accord with considerations of basic prudence at all. This is best demonstrated through a concrete example. The following case study of patient EVR, reported by Damasio et al. (Reference Damasio, Tranel and Damasio1990, pp. 91–92), vividly illustrates this:

EVR has a clear deficit in acting prudentially. Roskies (Reference Roskies2003) argues that vmPFC patients like EVR also exhibit a parallel failure to abide by certain norms of commonsense morality – such as those prohibiting breaking one’s promises or reneging on one’s responsibilities, etc. – to a degree sufficient to warrant the description “acquired sociopathy.” On these grounds, Roskies concludes that vmPFC patients suffer from a moral failure. However, since the performance of such patients on moral reasoning tests is (mostly) in the normal range, she infers that this failure consists not in a deficit in making appropriate moral judgments, but rather in a deficit in associating value with those judgments, so that the patients’ moral judgments have lost their normal motivational punch. In other words, vmPFC patients, according to Roskies, fail to be motivated to act in accord with the requirements of morality, despite (being capable of) knowing that what they are doing is wrong. In this way then, Roskies argues that empirical observations of brain-damaged patients support externalism about moral judgment. Moral judgments, by themselves, are not sufficient for generating moral motivation. Rather, they only move us when appropriately coupled to separate (and separable) motivational systems.

A number of philosophers have counters to Roskies’ arguments and alternate interpretations of these empirical observations. For example, Cholbi (Reference Cholbi2006) holds that vmPFC patients lack the moral beliefs that Roskies attributes to them. Consequently, he concludes, such patients pose no trouble for internalism. And Smith (Reference Smith and Sinnott-Armstrong2007) takes vmPFC patients’ failure to be motivated by their moral judgments only to show that their brain damage has rendered them systematically irrational. Leary (Reference Leary2017b) argues that vmPFC patients make “weaker,” or less confident, normative and moral judgments than normal subjects. As a result, she maintains, they are less motivated by these judgments than normal people – though, crucially, still somewhat motivated, which allows their decision making to be overruled by desires for greater or more immediate rewards. In these ways the internalist can seek to render her view consistent with the existence of neuropsychological patients with “acquired sociopathy.” This debate is ongoing and settled consensus on the significance, if any, of vmPFC patients for the internalist/externalist dispute has not yet been reached.

Much of the traditional philosophical debate concerning the internalism/externalism dispute, as we have seen, has turned upon the logical possibility of amoralists. In recent years, some have worried that the intuitions of philosophers engaged in this debate may have been “corrupted” by their theoretical commitments. Perhaps commitment to externalism affects whether one judges amoralists to be conceivable, and therefore logically possible, whereas commitment to internalism drives judgments of inconceivability. Hoping to make concrete progress in face of this seeming stalemate, experimental philosophers have decided to consult the intuitions of nonphilosophers – “the folk” – who almost certainly lack meta-ethical views and thus should be less subject to theoretical confirmation bias. Their reasoning here goes like this: If a substantial majority of nonphilosophers are ready to attribute moral judgments to moral subjects who lack the corresponding appropriate motivation, then the best explanation of this – ceteris paribus – is that “ordinary people” operate with an externalist conception of moral judgment and that the intuitions of internalist philosophers have been corrupted by their theoretical commitments. Likewise, for externalism and externalist philosophers, if the folk are unwilling to attribute moral judgments under such circumstances.

What do the data tell us? The first such experimental philosophy study (Nichols, Reference Nichols2002) sought to empirically investigate the popular conditional version of internalism, introduced earlier, according to which a moral judgment only causally suffices, by itself, for appropriate motivation in a rational agent, and an irrational agent can be wholly unmoved by her moral judgments. In order to test whether moral judgments necessarily produce corresponding motivations under conditions of full rationality, Nichols presented the following vignette to “philosophically unsophisticated undergraduates”:

After reading this scenario, subjects (N = 26) were asked whether John really understands that hurting others is morally wrong. Nichols’ results go like this: 85 percent of the subjects responded “Yes” and 15 percent responded “No.” These results suggest that ordinary people are mostly inclined to attribute moral understanding (and thus presumably moral beliefs) to John, despite his apparent rationality and lack of appropriate motivation. Nichols (Reference Nichols2002, Reference Nichols2004) takes this as evidence that the folk operate with some externalist conception of moral judgment, in which moral judgments can fail to produce corresponding motivation in a fully rational agent, one incompatible with the truth of conditional internalism (see also Strandberg & Björklund, Reference Strandberg and Björklund2013).

Nichols’ study, however, has been criticized on various grounds. First, as Joyce (Reference Joyce and Sinnot-Armstrong2008) observes, Nichols’ vignette does not make it explicit that John is practically rational. Given this, it is consistent with the data at hand that (at least) some of these subjects are operating with a conditional internalist conception of moral judgment, like the one advocated by Smith (Reference Smith1994), and simply believe that John is irrational. Having failed to control for this interpretation, Nichols’ study does not serve as evidence against conditional internalism. Second, Nichols’ vignette leaves too much implicit in other respects too. For example, it doesn’t rule out the possibility that John is somewhat appropriately motivated by his moral judgments, but this is overridden by stronger motivations to the contrary. And this is all that any plausible version of internalism entails. Third, the results of Nichols’ study have failed to replicate. Björnsson et al. (Reference Björnsson, Eriksson, Strandberg, Olidner and Björklund2015) ran a word-for-word duplicate of Nichols’ study, this time with 93 participants, and found that only 48 percent (rather than 85 percent) of subjects answered “Yes” to the question probing whether John understood that hurting others is morally wrong, with 52 percent denying that John understood this. Taken together, these criticisms suggest that further experimental inquiry is needed.

In response to this, Björnsson et al. (Reference Björnsson, Eriksson, Strandberg, Olidner and Björklund2015) report a number of studies that they interpret as suggesting that a majority of ordinary people operate with some internalist conception of moral judgment. The vignettes they use are far longer and more detailed and explicit than that used by Nichols (Reference Nichols2002). Consequently, they do not suffer from the problems described earlier. The most compelling result they report comes from a comparison of four different studies they conducted. All feature an agent Anna who can correctly classify actions as morally right or wrong but who reliably fails to act in accord with the requirements of morality. What varies between the studies is the explanation that is proffered for Anna’s actions. In one study (“Inner Struggle”), Anna was depicted as being motivated to act morally, but this motivation was then trumped by a stronger motivation to the contrary: to perform an action that she classified as wrong. Here 80 percent of subjects attributed Anna the moral belief that her action was wrong. In the second such study (“Listlessness”), Anna is presented as performing the same morally wrong action that she classifies as being wrong. However, here she is described as experiencing no motivation whatsoever to act in accord with judgment. But this is explained as being the result of her clinical depression, that has recently onset and left her listless and bereft of her previous zest for life, as well as her prosocial and other-regarding motivations. Here 70 percent of tested subjects were willing to attribute Anna the moral belief that her action was wrong. In the third study (“Psychopath”), Anna is again depicted as performing an action, that she classifies as morally wrong, and as experiencing no motivation whatsoever to refrain from doing it. However, here the explanation the vignette proffers for her failure to be so motivated is her clinical psychopathy. Now only 46 percent of test subjects attribute Anna the belief that action was morally wrong. Lastly, in the fourth such study (“No Reason”), Anna is again presented as doing something that she classifies as being morally wrong, but as having no motivation at all to abstain from so acting. Here, however, no explanation or reason is given for the absence of Anna’s expected motivation. In this case, a mere 36 percent of subjects are willing to attribute Anna the belief that her action was morally wrong.

What’s the significance of all this? Well, taken together, Björnsson et al. (Reference Björnsson, Eriksson, Strandberg, Olidner and Björklund2015) suggest that most tested subjects operate with an internalist conception of moral judgment: In the absence of a corresponding appropriate motivation, a rational agent cannot count as holding a moral belief. After all, the best explanation of the drop in “Yes” answers between the “Inner Struggle” (80 percent), where Anna has some motivation to act as morality requires, and “No Reason” (36 percent) or “Psychopath” (46 percent) conditions, in which she has no such motivation, is that most tested subjects regard moral judgments as entailing the existence of an appropriate motivation to act as they recommend in a rational agent. This is further supported by the fact that 70 percent of subjects were willing to attribute a moral belief to Anna in the “Listlessness” condition, in which Anna is presented as having no motivation to act in accord with morality due to her clinical depression, a mental state that is plausibly incompatible with full rationality.

However, there is still reason to be skeptical over whether experimental philosophy has settled the internalist/externalist dispute. First, results are not univocal over whether the folk operate with an internalist or an externalist conception of moral judgment (Björnsson et al., Reference Björnsson, Eriksson, Strandberg, Olidner and Björklund2015; Strandberg & Björklund, Reference Strandberg and Björklund2013). Second, and more importantly, the case has not yet been convincingly made that the folk’s conceptions of things – such as their conception of moral judgment – reliably carves at the joints of nature (Williamson, Reference Williamson2007). We have not been given good reason to think there is a straightforward link between what the folk think about internalism and externalism and the truth of these doctrines.

As in the internalism/externalism debate, empirical evidence has been marshaled by partisans in the cognitivism/noncognitivism debate. In recent years, there have been plenty of psychological and neuroimaging experiments investigating the relationship between making a moral judgment and experiencing certain emotions. On the face of it, if it can be demonstrated that (part of) what it is to make a moral judgment is to undergo a certain emotion (of admiration or anger, say), then that would constitute good evidence for noncognitivism about moral judgment – the view that moral judgments are (complexes of) noncognitive states, such as desires and feelings, and not beliefs. After all, perhaps, emotions are themselves noncognitive states, composed (at least, in part) out of paradigm noncognitive states such as desires, inclinations, and aversions, etc. (Prinz, Reference Prinz, Strandberg, Björklund, Björnsson, Eriksson and Olinder2015). Of course, any such evidence would be theory-laden: The fact, if it is a fact, that moral judgments are constituted by emotional states only counts as evidence for noncognitivism on the assumption that emotions themselves are noncognitive states. Cognitivists about emotion – such as Nussbaum (Reference Nussbaum2001) and Solomon (Reference Solomon1976) – would (clearly) reject this proposition. These philosophers hold that emotions are a species of cognitive state – for example, that what it is to experience guilt is to judge that you have engaged in a wrongdoing. Nevertheless, the dominant view in philosophical psychology is that emotions are noncognitive states (Prinz, Reference Prinz2007). Consequently, for the purposes of this discussion, we shall assume that noncognitivism about emotion is correct and thus that any evidence that moral judgments are constituted by emotions further counts as evidence for noncognitivism.

The evidence to be reviewed here in favor of noncognitivism also serves as evidence for the Humean theory of motivation. After all, the chief reason to disbelieve this Humean philosophical psychology comes from the case of moral motivation. Anti-Humeans contend that your moral beliefs – and your normative beliefs more generally – can motivate you to action in the absence of any desire. However, if moral and practical normative judgments are really noncognitive states, such as desires, as the noncognitivist contends, then even the case of moral motivation will be consistent with Humean psychology: Your moral judgments can motivate you because they are nothing but (complexes of) desires, sentiments, or emotions, etc. In which case, the only grounds for disbelieving the Humean theory of motivation will have been neutralized. So, although we shall present all the following evidence as pertaining to the debate between the cognitivist and the noncognitivist, the reader should bear in the mind that this evidence also speaks to the issue dividing Humeans and anti-Humeans about motivation.

There is now a significant amount of neuroimaging evidence suggesting that making a moral judgment cooccurs with experiencing an emotion. Indeed, every neuroimaging study investigating moral judgment seems to implicate brain areas known to be involved with emotion in moral cognition (Greene & Haidt, Reference Greene and Haidt2002; Prinz, Reference Prinz, Strandberg, Björklund, Björnsson, Eriksson and Olinder2015). For example, Heereken et al. (2003) instructed subjects to judge whether sentences are “morally incorrect” (such as “S steals R’s car”) or “semantically incorrect” (such as “S drinks the newspaper”). When subjects identified a sentence as being “morally incorrect,” their brains activated significantly more in areas associated with emotions relative to when they were identifying a sentence as “semantically incorrect.” Similarly, Moll et al. (Reference Moll, de Oliveira-Souza, Bramati and Grafman2002) had subjects make “right” or “wrong” classifications about so-called moral sentences, such as: “They hung up an innocent person,” and “factual sentences,” such as: “Stones are made of water.” Again, areas of the brain associated with emotions were significantly more active when subjects were making judgments about “moral sentences” relative to when they were making judgments about “factual sentences.” This conclusion – that areas of the brain underpinning emotions are implicated in tasks inducing moral cognition – is supported by a growing number of neuroimaging studies (Berthoz et al., Reference Berthoz, Artiges, Van De Moortele, Poline, Rouquette, Consoli and Martinot2002; Greene et al., Reference Greene, Sommerville, Nystrom, Darley and Cohen2001; Moll et al., Reference Moll, de Oliveira-Souza, Bramati and Grafman2002; Sanfey et al., Reference Sanfey, Rilling, Aronson, Nystrom and Cohen2003). The brain structures implicated by these studies include the anterior cingulate cortex, the insula, the orbitofrontal cortex, the temporal pole, and the medial frontal gyrus – all familiar players from emotion studies (Phan et al., Reference Phan, Wager, Taylor and Liberzon2002).

There is also a growing body of behavioral evidence that emotions have a causal influence on moral judgment. One prominent current in this literature is evidence that experiencing negative emotions looks to lead subjects to make more “morally critical” moral judgments than they would otherwise have made. For example, in one important study, Wheatley and Haidt (Reference Wheatley and Haidt2005) hypnotized subjects to feel a pang of disgust whenever they heard the neutral words “often” or “take.” Afterwards, they were asked to morally evaluate the protagonist of various stories, some of which contained one or other of these two trigger words. For example, they might hear about a congressman who “takes bribes” or “is often bribed.” Wheatley and Haidt found that the strength of subjects’ wrongness evaluations increases when the story contained one of these neutral trigger words, relative to evaluations of protagonists in morally equivalent scenarios. In other words, the experience of disgust, induced by a morally irrelevant trigger word, is observed to cause an increase in the strength of subjects’ moral denunciations. Furthermore, the effect remains when subjects are asked to evaluate the conduct of protagonists in morally neutral scenarios. For example, subjects report finding a student who is described as “often picking interesting topics in school discussions” – a morally neutral action – as morally suspect, even though they can’t explain why (“It just seems like he’s up to something…”). This result suggests that the experience of disgust, induced by a trigger word, can cause a subject to negatively morally evaluate a person who is described in ways that would not warrant such judgment. In a similar study, Schnall et al. (Reference Schnall, Haidt, Clore and Jordan2008) instructed subjects to morally evaluate the conduct of protagonists in described scenarios. For example, “Frank’s dog was killed by a car in front of his house. So he cut up the body and cooked it and ate it for dinner. How wrong was that?” Subjects were either sitting at a clean, tidy desk or a filthy, messy desk (featuring such disgust-inducing items as a used tissue, a greasy pizza box, a crusty drinking cup, and a chewed pencil, etc.) while performing this task. Schnall and colleagues found that subjects who were sat at the filthy desk morally evaluated the protagonists of the described scenarios more harshly than those who sat at the clean desk. Again this suggests that simply experiencing the emotion of disgust while reading the vignettes causes the subject to form more morally negative judgments of the described actors.Footnote ⁷

One natural explanation of these results is noncognitivism about moral judgment. For the noncognitivist, moral judgments cooccur in the mind with emotions, as the neuroimaging results described earlier suggest, because moral judgments just are constituted by certain (complexes of) noncognitive states – namely, the moral emotions of disapprobation, anger, admiration, guilt, etc. According to the noncognitivist, negative emotions induce more critical moral judgments, as the catalogued behavioral results described earlier indicate, precisely because such judgments are nothing over and above these emotions. In this way then, research in empirical moral psychology can be marshaled in favor of a noncognitivist philosophical psychology.

Of course, cognitivists about moral judgment push back here. They argue, for example, that these empirical results establish at most that there are causal relations between moral judgments and the emotions, and not that one is constituted by the other. For example, a cognitivist externalist who further holds that human subjects normally care deeply about morality can easily affirm that moral judgments cause emotions. After all, romantic music, gloomy weather, or high stakes sporting events – things many people care about – all cause emotions, without them (or our representations of them) being constituted by emotions. Given that, we should expect that our judgments about morality – something we generally care deeply about – should incite our passions. Indeed, there is empirical evidence that at least some moral judgments precede associated emotional experiences (Cusimano et al., Reference Cusimano, Thapa, Malle, Gunzelmann, Howes, Tenbrink and Davelaar2017), findings that are consistent with moral judgments causing said emotions.

But what about (negative) emotions causing moral judgments? How can the cognitivist explain this? One option is the experience of negative emotions draws our attention to morally relevant features of a situation (Prinz, Reference Prinz2007). This could certainly explain the results of Schnall et al. (Reference Schnall, Haidt, Clore and Jordan2008) and some of those reported by Wheatley and Haidt (Reference Wheatley and Haidt2005). However, it flounders in the face of Wheatley and Haidt’s finding, recorded in the same paper and briefly catalogued earlier in this section, that subjects who are hypnotized, such that they experience pangs of disgust upon hearing a neutral trigger word (“often”), negatively morally evaluate even protagonists who are not described as engaging in any morally wrong behaviors. There are no described morally wrong features at all in the scenarios in question. Consequently, it cannot be the case that these subjects are having their attention drawn, by their emotion, to features of the described situation that warrant moral condemnation: There are no such features. At first glance the cognitivist’s explanation appears inadequate.

However, cognitivists can explain these empirical results too. Very plausibly, the perception of a moral wrong, our cognitivist can hold, warrants not just the belief that a moral wrong occurred but also the (moral) emotion of disgust. One common way of expressing disapprobation toward a wrongdoer is to say something like “I’m disgusted by your actions, Jeremy.” Now, granting that we frequently experience (warranted) disgust upon learning of some morally wrong deed, and not when learning of morally neutral or praiseworthy actions, it follows that our experiencing disgust at someone’s action will be a reliable indicator, by our lights, that said action is morally wrong in one way or another. In this way then, the cognitivist can explain why a subject’s disgust at a described agent’s morally neutral action, unknowingly induced by a trigger word, can rationally lead her to form the belief that said agent is acting immorally.

This account also allows the cognitivist to explain the results of various “moral dumbfounding” experiments: experiments that show that people make moral judgments they cannot explain or rationalize. Haidt et al. (Reference Haidt, Björklund and Murphy2000) asked subjects to morally evaluate cases in which agents are described as engaging in harm-free acts such as consensual incestuous sex or cannibalism. One of the vignettes goes something like this: “Andy and Arianna are adult siblings who decide, after much reflection, to have consensual sex. They use contraception, both really enjoy it, and agree to keep it a secret. Did they do something morally wrong?” A total of 80 percent of subjects judged that their behavior was wrong, but they had great difficulty explaining why. Furthermore, subjects were presented with a decisive counterargument to any justification they gave. For example, some worried that they ran the risk of having deformed inbred children – but these subjects were then reminded that Andy and Arianna used contraception (or even knew themselves to be infertile). Others worried about effects on the community – but were then reminded that Andy and Arianna kept it a secret known only to themselves. Subjects tended to concede that these counterarguments were successful, but only 17 percent revised their moral judgment. The rest doubled down on their moral judgments and emotions: “Incest is plain nasty”; “It’s disgusting”; “It’s just wrong,” etc. Again this might be thought to constitute good evidence for noncognitivism about moral judgment: Our moral attitudes seem to bottom out in emotions of disgust, etc. However, the cognitivist can still explain this through appeal to the proposition that an agent’s experience of disgust at some behavior is a reliable indicator, by the lights of that agent, of its being morally wrong. This is consistent with both Andy and Arianna’s incestuous sex really being morally wrong and also with the subjects incorrectly believing it to be so. The confidence of cognitivists can also be bolstered by the fact that Royzman et al. (Reference Royzman, Kim and Leeman2015) conducted a study attempting to replicate the moral dumbfounding effect. They found that subjects who seemed to exhibit the moral dumbfounding effect nevertheless still tended to actually believe that the incestuous relationship would still cause harm to Andy and Arianna in the future and thus that their relationship was morally wrong on these grounds.Footnote ⁸ Furthermore, some subjects rejected the notion that wrongs entail harms, and rather held that incest was intrinsically or fundamentally morally wrong, in much the same way as harm is fundamentally morally wrong, and not wrong in virtue of some prior feature. Royzman et al. interpret their results as providing support for the cognitivist thesis that our moral attitudes bottom out in beliefs about morally pertinent features of things, and not in emotions of disgust, etc. As before, even though particular results may prima facie seem to support one position or the other, the empirical results can be argued to be consistent with the opposing position. Thus, the empirical work has not provided definitive evidence for either cognitivism or noncognitivism.

3.4 Concluding Remarks

This highly circumscribed review of the philosophy and cognitive science of moral motivation attempts to trace the convoluted lines of argumentation surrounding the major questions of internalism, Humeanism, and cognitivism. As we have illustrated, the nature of moral motivation is a “hinge issue” around which organizing debates in moral psychology and moral philosophy turn. Indeed, the disputes that we focused on – between the internalist and externalist, cognitivist and noncognitivists, Humean and anti-Humean – feature among the central issues in moral philosophy. There are other debates in moral psychology that connect to, or turn upon, the nature of moral motivation that we could not discuss here. For example, the nature of moral motivation matters to the debate over the possibility of altruism between the psychological egoist – who holds that all motivation to act is ultimately self-regarding – and the psychological altruist – who maintains that at least some motivation is ultimately other-regarding, or that we sometimes act for the sake of others as ends in themselves. We commend to the reader the other chapters in this collection (Chapters 12 and 13) which take up these and other matters.